Intra-national home bias: New evidence from the United States commodity flow survey

- We complement and extend Wolf's (2000) work on intra-national home bias for the US
- We estimate US home bias using all waves of the Commodity Flow Survey (1993-2012).
- PPML estimates point to a less relevant presence of border effects across US States.
- However, US internal market is less commercially integrated in 2012 than in 1993.
- Despite a lower presence of border effects across US States, those are increasing over time.

This paper estimates the US intra-national home bias in trade using Poisson Pseudo-Maximum Likelihood methodology to complement and extend Wolf's (2000) work. We use Wolf's data from the 1993 Commodity Flow Survey (CFS) and add the 1997, 2002, 2007, and 2012 waves. We claim that Wolf's home bias magnitude is overestimated due to the log-linearization of the gravity equation and the control for distance used in the cross-sectional study. Our results with panel data and latest econometric estimators show that the levels of US States home bias are in between 50% and 60% lower than in Wolf's. However, since 2002 the home bias has experienced substantial growth, which suggests that the US market is less commercially integrated in 2012 than in 1993.